13.6 PRECIPITATION VARIABILITY AND BARRIER-LAYER FORMATION IN THE NORTH INDIAN OCEAN

Shalini Mohleji and C.A. Clayson* Purdue University, West Lafayette, Indiana

1. INTRODUCTION whether these barriers truly act as barriers to deeper mixing and entrainment, and if so under what The Southwest Monsoon, occurring from the conditions. months of June through September, greatly affects and alters the air-sea exchanges in the Indian Ocean. 2. DATA, MODEL, AND METHODS The monsoon changes the wind and precipitation patterns over both the Arabian Sea and the Bay of Data used for initialization and forcing of the Bengal. The Arabian Sea experiences stronger model as well as validation come from several southwesterly winds and drier periods interspersed sources. Ocean and surface flux data for the western with intense short-lived precipitation events while the Pacific comes from the WHOI IMET buoy data Bay of Bengal encounters a decrease in winds but a (Anderson et al. 1996) during the TOGA COARE much greater amount of precipitation during the Intensive Observation Period (November 1992- monsoon season. As a result, these two areas February 1993). Surface flux data for the Indian undergo temperature and salinity variations in the Ocean region comes from the ECMWF (European ocean structure due to changes in the surface heat, Centre for Medium-Range Weather Forecasts) model moisture, and momentum fluxes associated with the analyses and TRMM (Tropical Rainfall Measuring monsoon (Sprintall and Tomczak, 1992). Mission) precipitation data. The time period of this The importance of salinity variability on the portion of the study is the monsoon season of May- thermodynamics and dynamics of the upper ocean September 2000. The TRMM precipitation data is the has gained an increased appreciation in the last few NASA DAAC daily 3B42 product, which is a 1°x1° years due to a phenomenon called the barrier layer. gridded product. The ECMWF model analyses used Much of the interest was initiated by the recognition of for all other surface fluxes have a resolution of the barrier layer feature in the tropical Pacific (e.g. roughly 1.25° x 1.25° and provides data four times a Lukas and Lindstrom 1991), and has more recently day. Both data sets are interpolated to 15 minutes been identified from climatological data in the Indian resolutions for forcing of the ocean model. Initial Ocean as well (e.g. Sprintall and Tomczak, 1992). temperature and salinity profiles for the Indian Ocean The barrier layer is a stable salinity-stratified but region come from the Levitus 1994 climatology. isothermal region at the bottom of the surface mixed A second-moment turbulence closure-based 1-D layer and is perceived to be an inhibitor of mixed layer model (Kantha and Clayson 1994) is entrainment due to its stability (Anderson et al., 1996). used to study the ocean surface and subsurface Barrier layer depths in the western Pacific are responses. The model includes the skin surface estimated to be on the order of 30 meters (Lukas and temperature parameterization developed by Wick Lindstrom, 1991). (1995), modified by Schluessel et al. (1997) to include Barrier layers have been found to affect not only the effects of precipitation. The model has a vertical mixing and entrainment, but also sea surface resolution of 1 m (maximum depth is 150 m) and a temperature (SST) (Anderson et al., 1996). The temporal resolution of 15 min. Temperature profiles theory is that when a barrier layer exists, mixing is have been validated over many time scales and in restricted to the near-surface water that is of nearly many locations (Kantha and Clayson, 1994); the same temperature, rather than the deeper cooler turbulence characteristics have also been validated water. With a lack of cooling but simultaneous (Clayson and Kantha, 1999). warming through solar radiation, the mixed layer warms up, effectively increasing SST as well. In the 3. METHODS western Pacific, isolated strong wind events such as westerly wind bursts are then thought to “punch The turbulent mixing, salinity, and temperature through” this stable layer and induce entrainment profiles are evaluated from model simulations as the cooling (Lukas and Lindstrom, 1991). model explicitly calculates turbulence properties in the This study investigates the variability in the water column. The validity of the model results creation of barrier layers due to monsoonal-induced presented in this talk rest on the ability of the ocean variations in the surface heat, moisture and model to accurately simulate variability in the momentum fluxes. We then further investigate temperature and salinity fields based on local surface ______heat flux observations. The model simulation shown *Corresponding author address: Carol Anne Clayson, here uses the western Pacific data. In order to Purdue University, Dept. of Earth and Atmos. Sci., West compare the simulation results and the buoy data, the Lafayette, IN 47907-1397; email: [email protected] buoy and model data are interpolated to every 0.5 m. Only values from the model simulations at the same 34.5 depths as the buoy data are retained before interpolation. The results shown in the talk will only 34.0 use the model data at the 1 m resolution. The barrier layer is identified as occurring when 33.5 the isothermal layer is deeper than the isohaline layer. 33.0 Mixed layer depths are calculated using the Anderson et al. (1996) methods. The isothermal layer depth is 32.5 6 NOV 20 NOV 3 DEC 17 DEC 31 DEC 14 JAN 28 JAN 11 FEB 25 FEB calculated by finding the point of maximum Fig 2. As in Fig. 1 but with sea surface salinities. temperature in the top 20 m and then calculating the differences downward until Dq = –0.030°C. The 0.00 isohaline layer depth is calculated by finding the point -10.0 of minimum salinity in the top 20 m and calculating -20.0 the differences downward until DS=0.0133 psu. -30.0

-40.0

4. MODEL VALIDATION Depth (m) -50.0

Sea surface temperatures and salinities from the -60.0

-70.0 buoy and the model simulations are shown in Fig. 1 6 NOV 20 NOV 3 DEC 17 DEC 31 DEC 14 JAN 28 JAN 11 FEB 25 FEB and 2, and and daily-maximum Fig 3. As in Fig 1. but with halocline depths. depths are shown in Fig. 3 and 4. The model simulations show reasonable variations when compared with the buoy data. The use of a one- 0.00 dimensional model precludes , which also -10.0 plays a role in the heat and salt balances in this -20.0 region. Observations during this time period showed -30.0 -40.0 that advection had been a minor factor in the upper Depth (m) -50.0 ocean heat content until roughly December 20 (Feng -60.0 et al. 1998). Similarly the only time period prior to the -70.0 December westerly wind burst in which the upper 6 NOV 20 NOV 3 DEC 17 DEC 31 DEC 14 JAN 28 JAN 11 FEB 25 FEB ocean salt content is strongly affected by advection is Fig 4. As in Fig. 1. but with thermocline depths. early November; advection contributes to a freshening of the upper ocean (Cronin and McPhaden, 1998). As 6. REFERENCES with the upper ocean heat content, the upper ocean Anderson, S.P., R.A. Weller, and R.B. Lukas, 1996: Surface salt content is strongly affected by advection during forcing and the mixed layer of the western the westerly wind burst (Feng et al. 1998). Anderson Pacific warm pool: observations and 1D model results. J. el al. (1996) also demonstrated that the salt and heat Climate, 9, 3056-3085. budgets were strongly affected by advection, Clayson, C.A., and L.H. Kantha, 1999: Turbulent kinetic entrainment, and errors in the prescribed fluxes energy and its dissipation rate in the equatorial mixed during the westerly wind burst in mid-December. layer. J. Phys. Oceanogr., 29, 2146-2166. Cronin, M.F. and M.J. McPhaden, 1998: Upper ocean 5. FURTHER REMARKS salinity balance in the western equatorial Pacific. J. Geophys. Res., 103, 27567-27587. Our presentation will focus on model results during Feng, M., P. Hacker, and R. Lukas, 1998: Upper ocean heat both the TOGA COARE IOP and the North Indian and salt balances in response to a westerly wind burst in Ocean during 2000. Results of the variability of the the western equatorial Pacific during TOGA COARE. J. barrier layer formation and its role in inhibiting mixing of Geophys. Res., 103, 10289-10311. will be presented. Kantha, L.H. and C.A. Clayson, 1994: An improved mixed layer model for geophysical applications. J. Geophys. Acknowledgments. We acknowledge with pleasure Res., 99, C12, 25235-25266. the support by the NASA TRMM program under grant Lukas, R. and E. Lindstrom, 1991: The mixed layer of the NAG5-9822. western equatorial Pacific Ocean. J. Geophys. Res., 96 33.0 (suppl.), 3343-3357. Schlussel, P., A.V. Soloviev and W.J. Emery, 1997: Cool 32.0 and fresh water skin of the ocean during rainfall. 31.0 Boundary-Layer Meteorol., 82, 437-472. Sprintall, J. and M. Tomczak, 1992: Evidence of the barrier 30.0 layer in the surface layer of the tropics. J. Geophys.

29.0 Res., 97, C5, 7305-7316. Wick, G., 1995: Evaluation of the variability and predictability 28.0 6 NOV 20 NOV 3 DEC 17 DEC 31 DEC 14 JAN 28 JAN 11 FEB 25 FEB of the bulk-skin SST difference with application to Fig 1. SSTs (buoy values in grey, model in black). satellite-measured SST. PhD thesis, University of Colorado-Boulder.